Tried policy with diffusion model

src/data/dataset.py

@@ -94,7 +94,7 @@ class NrvDataset(Dataset):
         # get indices of all 00:15 timestamps
         if self.full_day_skip:
             start_of_day_indices = dataframe[
-                dataframe["datetime"].dt.time != pd.Timestamp("00:15:00").time()
+                dataframe["datetime"].dt.time != pd.Timestamp("00:00:00").time()
             ].index
             skip_indices.extend(start_of_day_indices)
             skip_indices = list(set(skip_indices))

src/trainers/diffusion_trainer.py

@@ -48,6 +48,7 @@ class DiffusionTrainer:
         """
         return torch.randint(low=1, high=self.noise_steps, size=(n,))
 
+    
     def sample(self, model: DiffusionModel, n: int, inputs: torch.tensor):
         inputs = inputs.repeat(n, 1).to(self.device)
         model.eval()